Mixing in surround
- Thread starter Ryan Peters
- Start date
Illico
Samuel Le Tonquèze
I'm currently in the learning curve of ATMOS mixing with Cubase PRO and I'm in research how to deal with groups bus for bed mixing and how to deal with compression for loudness. For room reverb, you can use sends to a FX 7.1.2 track with a surround Reverb plugin (like Cinematic Rooms).
AudioBrewers
Are you listening?
We've been making videos in our youtube channel reg. mixing and "mastering" music for atmos. A question like yours landed just yesterday of a person asking how to 'master' for atmos.
The thing is, Atmos changes the rules of mixing from the very beginning, as there aren't really a fixed amount of baked channels from the moment of mixing, to the moment of delivering/listening = like in stereo.
The only golden rule is: stay on the loudness requirements (-18) and from then on, use your ears.
You can use your bed to mix all your music, if you pretty much desire to add mastering effects on the "master bus" like it's done in stereo, but generally speaking, because your material will be listened on several different setups, things like dynamics processors might give you more negative suprises than positive.
For atmos you have to think of Space, and how the energy of the sound is located around the listener, with that philosophy in mind, you could completely skip the bed (or use it for things that are more static, such as auxiliaries / reverbs) and dedicate yourself to mix with objects. And then mix 'master' each object independently to have it glue with your mix.
Finally, what I usually do, is creat a VCA track to link all my faders (bed + objects) to be able to sit my mix on the -18 loudness mark, but from then on, it's all about using your ears more than "this has to sound loud".
I gotta say, mixing for atmos is an incredible experience because it gives you too much freedom once you change your mentality towards how much will behave in a space, as opposed to a pair of speakers.
Finally, when it comes to DAWs, for development we use them all, and I cannot recommend Nuendo/Cub Pro enough, as it's the only DAW that supports three-dimensional objects AND ambisonics natively, meaning you can use things that have been developed with "immersive audio" in mind and the DAW will translate everything he way it should.
Finally, I compose and mix my music natively Atmos (completely skip stereo at this point, and just create a stereo mix from the Atmos render), so if you have any specific questions, do not hesitate to ask!
Some of the videos I've made you could find interesting:
(I will be putting this very session online for free next week for people to experiment)
As the last post will show, there is a lot of content on Atmos, but not as much on 5.1 I get it - it's old news. But there are definitely differences once a renderer and objects come inwith Atmos that 5.1 doesn't deal with. I guess the big takeaway I am hearing is - use your ears because there really aren't any rules and best practices other than keep music out of the center and be judicious when using the LFE.
charlieclouser
Senior Member
You might want to check out SAMP from HorNet plugins:
HoRNet SAMP virtual master bus for Dolby Atmos - HoRNet Plugins
Shape the Sound of Your Dolby Atmos Mixes with SAMP Audio Plugin. A Single Control Point for Mastering Your Mixes.
www.hornetplugins.com
It's an Atmos "virtual mix bus" with EQ + Comp + Limiter/Clipper. It's expensive though.... $18. Yes, EIGHTEEN dollars.
charlieclouser
Senior Member
I've posted at length previously about my knucklehead approach that dates back 20+ years, since before there was any surround support in Logic. Since Logic's surround features are still a bit lacking compared to ProTools and some other DAWs, I still use basically the same approach.
The good thing about Atmos taking over is that it can actually simplify the delivery process - as long as you're delivering to a dub stage where re-recording mixers will be doing all the Atmos stuff. They probably do not want the score delivered in anything resembling Atmos. Many mixers will actually prefer stereo-only stems, which allows them to spread things around in the Atmos field without having to un-bake multi-channel stems. If you are giving them more than just stereo stems, then quad is actually easier for them to deal with than true 5.1 or 7.1 stems. But that's a per-project question for the mixers on the stage.
The approach I've detailed elsewhere is basically this:
• I mix into 5.1 channel arrays per stem, but I leave the center channel and LFE channels empty, except in very special situations. The means I'm essentially mixing in quad - a "front pair" and a "back pair".
• Each stem has its own reverbs and delays, fed by sends from individual tracks. There is no "global" reverb. I use stereo-only reverbs and delays, and use slightly different settings on the effects that feed the rear pairs so that those reverbs are a little longer, with less early reflections, to add more distance to the back. Occasionally I've used matched sets of stereo impulse responses taken from the front / rear of the source room to emulate a quad reverb.
• The stems are summed within the DAW to also create a composite mix. There is no processing at all on the sub-master for that composite mix - it is just a "straight-fader" sum of the stems. This channel array is what goes to my speakers and meters.
• Each stem has its own compressor / eq / limiter inside my DAW. I adjust the ceiling on those limiters to prevent the composite mix from clipping.
Lately, however, I've been delivering for projects that will be mixed in Atmos and the mixers told me they'd prefer it if I just gave them stereo stems - but more of them. Since my DAW can feed 64 channels to my print rig, what used to be ten 5.1 stems can now be 32 stereo stems, or 16 quad stems. This is a lot more flexible and simple for the mixers to deal with since they're not trying to figure out how to link a center channel with its accompanying L+R channels in an Atmos array. On some projects I still print in quad, but the rear pair is basically a wet-only pair. Then the mixers can use that pair or not as needed.
Thanks for sharing!
May I ask please what are your plugin preferences for compression and limiting on those 5.1 buses/stems?
AudioBrewers
Are you listening?
There is a huge difference between atmos and good olf 5.1:
- 5.1 is, again, baked channels, that expect a listener to have a 5.1 array.
- Atmos has no defined channels. and instead, there are 128 "elements" that are decoded on the spot depending on the output array.
It might sound like Atmos is an extension of 5.1 - but in reality, 5.1 is an extension of Stereo (baked channels), and Atmos is a completely separate thing.
Warning regarding the LFE channel: When decoding from Atmos to 2.0, the LFE channel WILL BE DISCARDED by the decoder (as per Dolby's guidelines), so in music, I would advise against using the LFE channel, and if you are, do it with caution, always being aware that your mix should work without the LFE present.
So in a way, never mix 5.1 thinking it can become Atmos, although the contrary (mixing in Atmos, exporting in 5.1) would work just fine.
Recently, every time I do something, I just do it directly in Atmos, and then just export whatever I need (stereo, 5.1 for DCPs, or the Atmos ADM file), it just works better.
charlieclouser
Senior Member
• In the olden days I'd use TC MasterX5 on each stem - that's an Ozone-style "mastering processor" that's broadly similar to the TC Finalizer hardware unit. Very powerful upward-compression that doesn't give you totally clipped, squared-off waveforms. Very natural sounding, and with 5 bands the subs get hit separately to the pitched low end. I tweaked the crossover points between the bands to give me: Subs (below 60hz), Bass (60-500hz), Honk (500-1.5kHz), Clank (1.5-4kHz), and Fizz (4k and up). But that software ran on the now-discontinued PowerCore DSP platform, so it's dead to me now. TC / Behringer does make a native plugin that they call MasterX, and it's good... but not as good.
• Most of the time I use Waves L3-LL MultiMaximizer as a per-stem limiter. It's very transparent and puts an absolute brick wall on the levels. It is five bands, like my beloved MasterX5, but it does exhibit some squared-off waveforms as the result. In a shootout, this was the closest to the sound of my old MasterX5, despite the waveforms that can look clipped. The great thing about L3-LL is that the "LL" stands for "low latency", so I can leave it in place at all times and it doesn't mess with the feel of playing live MIDI. I haven't measured the latency but it's very short.
• Sometimes I will use Ozone, or just the Ozone Maximizer. It's more DSP hungry than most others, but it can pack a wallop and is very flexible.
• For really aggressive cues I will often use Waves Abbey Road Saturator. It can get fierce, so it takes a delicate hand to get it not to grind and distort, but it can light a fire under a stem for sure. Typical uses would be on a drums stem and / or on a strings stem that has lots of low spiccato parts that need more bite to avoid turning into mushy mud.
• Other than a stock Logic EQ set to high-pass at 20hz on every stem, I tend not to use an EQ on the stems except on special situations. When I do I favor a vintage Neve-style EQ, either UAD 1073 Channel Strip or just Logic's Vintage Console EQ (which sounds great). Sometimes I will put a digital-style EQ on a stem to cure some problem or other, and this is usually FabFilter Pro-Q3. But most of the time my per-stem EQs are bypassed, and all corrective EQ is done per-track.
Thanks! I appreciate that.
I am too an orphan of MasterX (3 though, I never had 5) on the Powercore. It used to be on my 2bus and on vocals. I loved it for the powerful multiband control too. (Dunno why they never came up with a 1:1 100% identical version for native systems...)
I'm a fan of L3 and it was my limiter of choice until I got ProL2.
I will definitely try the LL version on the stems buses.
Will also give a try to AR Saturator.
Maybe it is the only Abbey Road plugin I don't own. As a Beatles fanatic I automatically bought anything Abbey Road but I wasn't fond of any of the latest releases from the series.
Plates, Chambers, TG mastering chain and especially the RS124 were a total let down for me.
(Love the older Waves REDD, TG12345 and ADT and I love the Rs127 found in the Softube AR Brilliance pack so much that I managed to find a hand built HW clone of it).
Until now I have been handing 5.1 mixes to an external engineer and I'm currently in the position of finally delivering my first 5.1 mixes/stems directly to the dub stage. So I'm kind of finding my feet in this and developing a workflow.
Thanks again!
Olympum
Active Member
This thread and other recent ones on mixing on quad got me curious to setup two older monitors as Ls/Rs. The experience is amazing, and it’s really pleasant to compose and make music in quad. I am partially deaf on my right ear, which causes me to struggle while mixing, and specially hearing “space” and reverb in stereo. Whereas in quad I am finally noticing how wet everything was and I can almost position it.
It’s really interesting to see the approach others are following for mixing in surround. Initially I went the route of converting my project template from stereo to 5.1 surround in which, like Charlie, I only used the front and rear speakers and left Lfe/C empty. I routed the close and main mics to the front, the surround mics to the rear, and everything through a CRP 4.0 reverb instance (to avoid the Lfe/C). The sound is … wow. Sure it’s not balanced, not the same speakers, but I don’t care. It sounds great and it feels more musical.
And then I hit Atmos … or rather Atmos hit me.
I wanted to see how to export from the 5.1 setup into an Atmos mix. The things is that I have not seen any best practices yet for mixing music for Atmos (some exist for film, dialogue, effects, music, but not for just music releases in Atmos), whether going to Atmos from stereo or from 5.1. I tried the usual suspects, from just routing the 5.1 to an Atmos bed, to splitting the signals into stereo pairs and using only objects. I have finally ended up in a hybrid configuration with a bed with reverbs, and objects for close, main and room mics stereo pairs.
I setup it up in VEPro with multiple stereo pair audio returns: one for main mics, one for room/surround mics, and then a list of close mics per instrument section, in my case WW, horns, trumpets, etc. In the DAW (Cubase, but I am sure this is true for all), all the audio returns feed into objects in the Atmos mix. For close mics, I narrow the stereo a bit, and put them far in front of the user, with some minor rotation. I place the mains and room mic mixes also on front of the listener, but the main mic a bit closer, and the room mic very close to the listener, and higher up. Putting things behind the listener is super cool, but not very realistic, at least in my opinion. Finally, I put a send on the close and main returns feeding a CRP 7.1.2 instance going into the Atmos bed. Sometimes I’ll add a touch of the reverb send too to the main mic for a library that is dryer, or if for example I only have mid mics.
Overall this allows me to mix and match libraries. I start with the main mic and bring it up to an equal level. Then add the room mic from the bottom. And finally bring it closer using the close. I then add the reverb send levels to glue it up. With automation I can focus on instruments by varying the reverb and close gain. I guess I could also bring the corresponding object closer to the listener, but I think that’s a bit too much.
This is a surprisingly flexible as a setup. It’s super easy to change and make it dryer or wetter, and to quickly reposition things around and sculpt the sound to be what I want it to be. I went through multiple wet libraries where I have individual mics (all in their own rooms), and managed to make them “fit” quite well in my “own room.”
Even though I originally I would only work with Atmos for mixing, exporting stereo wav files and then moving them onto an Atmos mixing session, I have now setup my writing template in Atmos too. I have found that in my M1 Max has plenty of CPU, as my instruments are off loaded to a VEPro server. The Dolby Renderer is not very heavy, nor introduces significantly amount of latency. Maybe it’s because I don’t mind working at 512 bytes buffer as required for Atmos, and I usually play ahead of the beat anyway.
This setup has brought a new life into my libraries, as I can finally hear things properly (the good and the bad things about them too).
I kind of came to my setup by trial and error, but still wonder what other “best practices” others are following for mixing music in surround, particularly Atmos.
It’s really interesting to see the approach others are following for mixing in surround. Initially I went the route of converting my project template from stereo to 5.1 surround in which, like Charlie, I only used the front and rear speakers and left Lfe/C empty. I routed the close and main mics to the front, the surround mics to the rear, and everything through a CRP 4.0 reverb instance (to avoid the Lfe/C). The sound is … wow. Sure it’s not balanced, not the same speakers, but I don’t care. It sounds great and it feels more musical.
And then I hit Atmos … or rather Atmos hit me.
I wanted to see how to export from the 5.1 setup into an Atmos mix. The things is that I have not seen any best practices yet for mixing music for Atmos (some exist for film, dialogue, effects, music, but not for just music releases in Atmos), whether going to Atmos from stereo or from 5.1. I tried the usual suspects, from just routing the 5.1 to an Atmos bed, to splitting the signals into stereo pairs and using only objects. I have finally ended up in a hybrid configuration with a bed with reverbs, and objects for close, main and room mics stereo pairs.
I setup it up in VEPro with multiple stereo pair audio returns: one for main mics, one for room/surround mics, and then a list of close mics per instrument section, in my case WW, horns, trumpets, etc. In the DAW (Cubase, but I am sure this is true for all), all the audio returns feed into objects in the Atmos mix. For close mics, I narrow the stereo a bit, and put them far in front of the user, with some minor rotation. I place the mains and room mic mixes also on front of the listener, but the main mic a bit closer, and the room mic very close to the listener, and higher up. Putting things behind the listener is super cool, but not very realistic, at least in my opinion. Finally, I put a send on the close and main returns feeding a CRP 7.1.2 instance going into the Atmos bed. Sometimes I’ll add a touch of the reverb send too to the main mic for a library that is dryer, or if for example I only have mid mics.
Overall this allows me to mix and match libraries. I start with the main mic and bring it up to an equal level. Then add the room mic from the bottom. And finally bring it closer using the close. I then add the reverb send levels to glue it up. With automation I can focus on instruments by varying the reverb and close gain. I guess I could also bring the corresponding object closer to the listener, but I think that’s a bit too much.
This is a surprisingly flexible as a setup. It’s super easy to change and make it dryer or wetter, and to quickly reposition things around and sculpt the sound to be what I want it to be. I went through multiple wet libraries where I have individual mics (all in their own rooms), and managed to make them “fit” quite well in my “own room.”
Even though I originally I would only work with Atmos for mixing, exporting stereo wav files and then moving them onto an Atmos mixing session, I have now setup my writing template in Atmos too. I have found that in my M1 Max has plenty of CPU, as my instruments are off loaded to a VEPro server. The Dolby Renderer is not very heavy, nor introduces significantly amount of latency. Maybe it’s because I don’t mind working at 512 bytes buffer as required for Atmos, and I usually play ahead of the beat anyway.
This setup has brought a new life into my libraries, as I can finally hear things properly (the good and the bad things about them too).
I kind of came to my setup by trial and error, but still wonder what other “best practices” others are following for mixing music in surround, particularly Atmos.
Last edited:
quickbrownf0x
I had fun once. It was awful.
Man, that's a lot of cash. Good thing I just started my weekend paper route.You might want to check out SAMP from HorNet plugins:
HoRNet SAMP virtual master bus for Dolby Atmos - HoRNet Plugins
Shape the Sound of Your Dolby Atmos Mixes with SAMP Audio Plugin. A Single Control Point for Mastering Your Mixes.www.hornetplugins.com
It's an Atmos "virtual mix bus" with EQ + Comp + Limiter/Clipper. It's expensive though.... $18. Yes, EIGHTEEN dollars.
